When the ground trembles

An assessment of earthquakes and the losses they cause

Recent decades have seen a significant rise in losses due to earthquake disasters. For the insurance sector, it is therefore of growing importance that earthquake risks - and, in particular, the potential losses - be assessed as precisely as possible.

History presents terrifying examples of catastrophic earthquakes. There are only some of the most famous events:

• —
  2005: (Pakistan/India), magnitude 7.6, 88,000 fatalities
• —
  2004: Northern Sumatra (Indonesia), Sea quake with tsunami, magnitude 8.9 — more than 200,000 fatalities
• —
  2003: Bam (Iran), magnitude 7.9 — 26,000 fatalities
• —
  1976: Tangshan (China); magnitude 7.9 — 242,000 fatalities
• —
  1923: Tokio (Japan), magnitude 7.9 — 140,000 fatalities
• —
  1906: San Francisco (USA), magnitude 7.8 — at the time the most expensive natural catastrophe

Recent years show that such catastrophes can happen again at any time. To make things worse, due to rapid and dense urbanisation, the scenarios of the future may be even more horrifying. The cities along the circum-Pacific "ring of fire" are classed as highly exposed.

Immense losses are foreseeable for people, economies, and insurers. But the exposure potential from quakes within continents, too, or along the Mediterranean/trans-Asian earthquake belt should certainly not be underrated.

More quakes mean a greater potential for property damage

The earthquakes at Northridge, 1994 (insured loss at the time: US$ 15.5 bn) and Kobe, Japan, 1995 (economic loss: > US$ 100 bn; insured loss: US$ 3 bn) were the most serious in the industrialised world for many decades. With their extraordinarily high property losses, coupled with a relatively low number of casualties, they are typical of a trend that has been observed: as the economic development of a region progresses, there is a shift from loss of life to property damage.

More development leads to more global interdependence

Advances in the development and industrialisation of a region are also associated with growing international interdependence. This means that the scale of natural catastrophes is greater than ever due to the indirect losses. A devastating quake in the Tokyo region could now generate losses amounting to between US$ 1,000bn and 2,000bn - staggering dimensions that would put the squeeze not only on every single national economy but on the entire financial world.

But how probable are such scenarios?

Although the magnitude and location of such a scenario may be quite assessable, the timing of a potential quake is not. Given the present state of research, it is only possible to make statistical statements on the occurrence probability.

In the case of San Francisco, for example, there is a 60% probability of an earthquake with a magnitude of 6.7 or above occurring in the next 30 years. Short-term predictions for the immediate future (e.g. a day or a week), such as might allow precautions - like evacuating a city - to be taken in good time, are not possible at present, and this is unlikely to change in the future.

Consequences for the insurance industry

Nonetheless, insurers are now in a better position to assess their own loss potential on the basis of scientific data and information on the assets insured. Computer models simulate thousands of quakes and calculate the possible losses, taking any secondary effects (like landslides, liquefaction, fires after earthquakes, and tsunamis) into account.

On this basis, we can calculate premiums that are commensurate with the risk, even though this still involves some major uncertainties. Just as important for the insurance sector is an assessment of the size of rare losses, such as those that occur every 500 or 1,000 years, so that it can limit the risk it assumes to an amount it can actually afford to pay. The monitoring of this so-called accumulation is a central task for reinsurers.